Compressing/Mixing to be more studio-like

Really talented bands (or at least ones who are obsessive over the mixing) generally release their songs where most notes peak the channels, without overflowing. A really good example is any song from Linkin Park’s Meteora album, if you have a copy of it to rip to Flac.

I’ve been given the task of fixing some Vinyl rips. After equalizing it with the FFT filter just right, the last step is to compress it and give it the oomf and power it needs. Some despise that and call it the “loudness war”, but there are times when it works just right for songs. The issue is that I don’t know how to max each hard note to max the channels like a perfect flac copy of the song would if it were released in this decade.

After opening the Compressor dialog, -10dB was the best threshold I could get to amplify. I’m not sure exactly what Noise Floor was, but I tested it at both extremes (-80dB and -20dB) and it didn’t really make a difference. I set the ratio to 10:1, the max. Attack time was 0.1 seconds since the amplification needs to be immediate, and decay time was also immediate at 1.0 seconds because it needs to be sharp and quick. Lastly, Compress based on peaks was checked, and I clicked OK, and the notes again didn’t generally hit the peaks. When I increased the threshold to allow lower-contrasted notes in, the same applied, and the dynamic contrast on top of that became really, really bad (“loudness war”). I quit fiddling with it after about 15 minutes and called it quits.

I’ve always been disappointed with Audacity’s lack in real life application, so I’m looking forward to tips here and there to get it comparable to proprietary software—you know, without violating my four freedoms. :mrgreen:

It’s all about using the right tools for the job.
Since you’re concerned about freedom I presume that you are using Linux. Audacity is a great tool for editing and for quick and efficient mixing, but it’s not really a mastering tool. Have a look at “Ardour 2” and “Jamin” http://jamin.sourceforge.net/en/about.html

That is an inconvenience since I use Fedora 12 on my laptop and Vista SP2 on my desktop where I do my real stuff and haven’t gotten around to dual-booting. I’ll give those a try, however, so thanks. :slight_smile:

<<>>

Have you messed around with Chris’s Compressor?

Chris’s Compressor
http://pdf23ds.net/software/dynamic-compressor/

If you’re hand-tuning the Audacity compressors note by note, this may be a revelation to you.

Koz

You’re all worthless :stuck_out_tongue: The Hard Limiter was exactly what I was looking for. I know what each slider does and I use it all the time now. :slight_smile:

Hopefully at least one of you learned something new.

<<>>

Then you win. Hard Limiting is a special purpose effect – like a fuzz guitar – that can do serious damage to the sound. That’s why it’s not right at the top of our recommendations.

Koz

JAMin features:

  • Linear filters
  • JACK I/O
  • 30 band graphic EQ
  • 1023 band hand drawn EQ with parametric controls
  • Spectrum analyser
  • 3 band peak compressor
  • Lookahead brickwall limiter
  • Multiband stereo processing
  • Presets and scenes
  • Loudness maximiser

There’s no “Professional Audio Filter” (regardless of what this topic suggest :smiley:)
As I said before, it’s about using the right tool for the job, and often the job will require several tools.
Exactly which tool to use, how to use it and when to use it depends on the materials and the desired result.
JAMin offers a comprehensive set of tools for performing the task that you asked about, but if you’re happy with the hard limiter then JAMin is probably more than you need. :wink:

here are two of the things i mentioned elsewhere

third item

did not find all these in the official list of other stuff

now i need to find a way to capture all these easily
and put in one place for you all to update the official list
or make an unofficial list or whatever

I’ll start by writing a simple web page for posting. You can start in NotePad or TextEdit and give them to me in PM as three text lines.

General Description. Comments. Platform(s)
Title of the Tool
Web Address.

General Description. Comments. Platform(s)
Title of the Tool
Web Address.

General Description. Comments. Platform(s)
Title of the Tool
Web Address.

I may not post it in that order. I haven’t thought about it that far ahead. I also need to give a nod to ease of updating.

Sound Compressor. Volume Compressor. All Platforms
Chris’s Compressor
http://pdf23ds.net/software/dynamic-compressor/

I don’t want it to get too deep too fast. We’ll never be able to find anything.


Chris’s Compressor
All Platforms
Sound Compressor. Volume Compressor.
http://pdf23ds.net/software/dynamic-compressor/

How about major divisions by platform? Chris would appear three times.

— Microsoft Windows —
Chris’s Compressor
Sound Compressor. Volume Compressor.
http://pdf23ds.net/software/dynamic-compressor/

Koz

I think it would be a good idea to have a sticky topic on the forum (as suggested by Gale in this post https://forum.audacityteam.org/t/listing-useful-programs-other-than-audacity/14317/7 ) as this would allow others to add suggestions. I realise there is a risk of it becoming a “spam magnet”, but I think it is worth giving it a trial run, especially as whomper appears to be keen to add some software suggestions. As previously suggested, if the topic becomes popular, then we can look at devising a work-flow for transfering suggestions to the appropriate lists on a routine basis.

I have started a sticky topic here: http://forum.audacityteam.org/viewtopic.php?f=28&t=31394&p=86913#p86913
If any of the moderators would like to add or change anything in the starter post, please feel free to edit it.

I can see how sample precision lowers when you decrease the Wet level. Can you be a little more specific on the “damage” done, exactly?

Hard limiting is basically the same effect as distortion that is caused by recording too loud, though it can be made less harsh than digital clipping by increasing the “residual” level. Hard limiting distorts the peaks by squashing the tops of the waveforms, thus creating significant amounts of odd harmonics. The effect is very much like a fuzz box, except that the distortion is limited to the high peaks, whereas a fuzz box will be set with a much lower threshold so as to create distortion on lower level sounds.

To get a clear idea of what hard limiting does, try setting it to an extreme setting of “-20, 1, 0”

I already know exactly what Hard Limiter does. When there are really high sharp peaks that prevent the rest of the music from reaching an appropriate level, they need to be lowered so the track can be amplified without the peaks clipping. Hard Limiter sounds like it’s exactly the tool to do that. It sounds like it could even the peaks so that through the whole song they are able to touch the edges of each channel. So to me, it sounds like audible damage can only be done if you lower the Wet level drastically, which could be compared to amplifying to -50dB and then reamplifying. But on the other hand, Wet level needs to be lowered enough to anti-alias the peak clipping. But I don’t think that’s a problem at all if done in the right amount since all that can be done is a re-interpolation: you won’t be able to measure a difference in precision or notice that there is any noise introduced.

The way that the hard limiter works is that it clips the peaks, though it can be set to do so more gently than hacking off the peaks flat. The harmonic distortion is of an identical kind to clipping the peaks, but not as much.

The hard limiter splits the sound into three components - the upper peaks (the bits above the threshold), the negative peaks (same, but negative going), and the middle bit.

If you set the “residual level” to zero, then what you get is just the middle bit. This is identical to clipping the peaks.
Here is a close-up of a sine wave that has had the hard limiter applied with settings of -6, 1, 0
firsttrack000.png
The “Residual” part, is the other two parts. We can isolate these two parts by setting the Wet level to zero and the residual level to 1
firsttrack001.png
Adjusting the “Wet” level scales the amount of clipped audio (the middle section). When set to 0, it is scaled down to silence, when at 1 it is at the original volume (first image) and when at 0.5 it is scaled (amplified) to half volume.

Adjusting the “residual” level scales the “peak” component in the same way.

Usually you would want to keep the Wet level at 100% and then add in a proportion of the “residual” components to “soften” the clipping effect.
Setting the residual level to about 0.3 is usually a good compromise, and produces a result like this:
firsttrack002.png
If you look at the spectrum (Analyze menu > Plot Spectrum), you will notice that soft clipping produces the same odd harmonics as hard clipping, but at a lower level.

There are other types of limiters that produce lower distortion levels, but this type is simple and quick to use, and when used to just clip occasional peaks it is quite adequate. The distortion can in some situations be quite effective, for example, if used on a slightly dull recording of a cymbal or snare drum, it not only limits the peaks, but the distortion can add a bit of extra brightness to the initial hit. Odd harmonics are often described as less musical sounding than even harmonics, but in the case of percussive instruments such as these it is less important as the sound is essentially inharmonic.

Steve, you just posted what I know exactly already.

You mentioned “The way that the hard limiter works is that it clips the peaks, though it can be set to do so more gently than hacking off the peaks flat.”. This is exactly what I want, isn’t it?

dB limit sets the cutoff threshold. Residue level sets what percent volume the cutoff area should be, where 0 means 0% and 1 means 100%. Wet level means feathering the cutoff area, where 1 means feather only as far as -1 and +1 (meaning, no feathering at all) and 0 means feather all the way to 0 (which makes the track have no volume).

Tell me something I don’t know like the negative effects.

Edit: I just read your last paragraph and it’s informative. :smiley: I’ll look into the spectrum myself.

Maybe I missed something, but why compress? And how do you mix to make it more studio?

In what manner was your original copy recorded? Equipment wise.

Why do you need to make it more studio? Is it too flat? Or poppy or lispee?

And is it stereo?

Do you just want it to have more punch?

The reason we all like Chris’s Compressor so much is that it increases the low passages of the show and suppresses the loudest ones automatically without seeming to do anything.

It’s directly comparable to the “invisible” volume compression that an FM radio station does. I don’t have it around any more, but I used to have A/B comparison test between the “internet” version of “Car Talk” and the version on LA’s KPCC Radio. Millions of people are happy listening on the radio and compared to that the internet version is almost unlistenable. They just posted the raw feed and every time Tom would laugh, the volume increase was enough to set off smoke alarms and scare the cat.

A quick trip through Chris and the volume variations are back to where they should be for a comfortable listening experience.

You could not get that with straight peak compression.

As as I said, if you found a tool that works for you, then you win. The peak lopper-offer is much more indicated if you have a noisy phonograph record you’re trying to process. Great differences between the pops and the music. Chris is helpless there.

Koz

By the way, you’re speaking the advertising words of CBS Laboratories, the designer of the Audimax and Volumax FM Radio compressors. They reigned broadcasting for years. The Audimax was the slow, sloppy overall volume setter and the Volumax (or Volume-Ax in local speak) would come up behind it and surgically whack off the offending peaks.

It was OK at best, but it was the state of the art for a while.

Koz